MoneyLIVE North America: Powerful Panel Perspectives
By John San Filippo
MoneyLIVE North America was held in Chicago September 14-15, 2026. For the second consecutive year, Finopotamus was privileged to be an official media partner for this growing event.
Conference sessions were presented as a mix of fireside chats, keynote addresses and panel discussions. The panel discussions brought together a diverse range of financial industry practitioners, each with unique insights. Finopotamus attended two of these panel discussions.
Across both discussions, participants repeatedly converged on the reality that AI maturity is no longer measured by isolated proofs of concept, but by foundational data architecture and disciplined execution. Rather than debating the theoretical promise of frontier models, panelists focused on the unglamorous mechanics required to make AI functional: semantic data layers, standardized ontologies, and real-time data pipelines.
A second unifying thread was the balance between speed and control. Whether managing enterprise risk frameworks or navigating marketing compliance, speakers emphasized that institutions should not try to reinvent frontier models themselves but instead assemble strong guardrails that allow them to operationalize third-party capabilities safely.
Finally, both discussions warned that customer loyalty will not be won through generic automation, but through consistent, cross-channel experiences that treat the consumer as an individual rather than a sales target.
Here is a closer look at the two panel sessions.
The AI-Powered Bank: Unlocking Enterprise Agility and New Revenue Models
Moderated by Rahul Dubey, Chief Growth Officer at Publicis Sapient, this panel featured Kelley Conway, Chief Data and Analytics Officer at Northern Trust; Greg Clark, Managing Director for AI and Advanced Analytics at Bank of Montreal; and Mike Blanco, Managing Director of Emerging Technology Risk and Governance at Midwood Partners. The discussion focused on transitioning AI from internal cost-cutting into sustainable revenue generation.
Conway noted that while financial institutions have established clear returns on investment in risk reduction and contact center productivity, driving top-line revenue remains the industry’s central hurdle.
“AI has clearly moved beyond experimentation at this point and POCs, and we are really getting this into production,” Conway said. “Productivity, faster decisions, that is where this has all started. Now we are on the verge of revenue, and how to use the right models for the right use cases and optimize those tokens and really kind of take that next scale.”
Conway stressed that reliable outputs depend entirely on data structure rather than model size, highlighting Northern Trust’s internal testing with semantic layers.
“AI will do its damnedest to give you an answer every single time, whether it knows the answer or not,” Conway stated. “And the only time that it gave a consistently deterministic accurate answer was when it was in the semantic layer and we had that metric defined.”
Blanco addressed the ongoing debate regarding whether banks should build their own foundational models, urging institutions to stick to their traditional strengths. “What they need to let go of is just understanding that they are not a technology company,” he said. “They should not be competing with Google, Anthropic, or OpenAI. They need to focus on their core competencies and their competitive advantage and continue to excel.”
Blanco also noted that standard risk frameworks still apply to emerging technology. “Risk management is risk management is risk management,” he observed. “Whether it’s AI, whether it’s digital assets, it’s all about just identifying the risks, building that back into your taxonomy, identifying the failure nodes, and asking yourself, does this fit within my risk appetite?”
Clark, drawing on his background as a research mathematician, pointed out that the definition of institutional trust is expanding as organizations navigate both small and large language models.
“Business moves at the speed of trust,” Clark explained. “At BMO, whenever we are thinking about how do we not only continue to move with the pace and scale, but do so in a safe and effective way, it all comes back to trust in a variety of ways, whether it is how we trust the system, how we trust the model, how we trust one another, and how our clients ultimately trust us.”
Marketing in the Age of AI: How Should Bank Strategies Adapt?
Moderated by Juliette Foster, this session brought together Rahul Rajendra Prasad, Senior Vice President of Technology at Citizens Bank; Dhaval Gala, Director of Campaign Analytics and Customer Interactions at Scotiabank; and Luis Landivar, Head of Solutions Consulting at Naehas. The conversation explored how banks can deliver hyper-personalized marketing without drowning in unverified content or violating regulatory mandates.
Prasad outlined how Citizens Bank approaches the shift from traditional volume-based promotional blasts to intent-driven engagement.
“We have typically looked at marketing as conversion rates and targeting campaigns at customers,” Prasad noted. “Now we are starting to look at how do we get closer to the customer and meet the customer where they are in their financial journeys and start to really nudge them and be able to provide meaningful outcomes for them.”
Prasad described a five-stage framework for agentic decisioning: know me, see me, understand me, guide me, and learn about me. He emphasized that marketing must feel uniform across every touchpoint.
“How seamless can we make the banking experience itself?” Prasad asked. “When a recommendation goes from a marketing campaign, how is that a consistent message across branches when they walk into a branch and talk with an engagement manager, if they call into a contact center, or they log into their mobile app and their online banking experience?”
Gala highlighted the operational friction that occurs when generative tools produce dozens of campaign variations faster than compliance teams can inspect them.
“Sometimes the content AI and the decisioning AI do not talk to each other,” Gala explained. “What happens is you will have decisioning that is happening at scale, but you have all these content fragments and content versions. And marrying that in a cohesive ecosystem, in a cohesive system that is closed loop, that is learning, that is a big problem.”
Gala added that certain high-stakes decisions must remain off-limits to autonomous agents. “Pricing decisions, offer decisions, marketing strategy, those are going to be driven by humans,” Gala stated. “Things that are client-facing, customer-facing, will always have human in the loop.”
Landivar advised institutions to re-engineer their compliance workflows by separating fixed regulatory elements from flexible marketing copy before running automated generation.
“The competitive advantage is really not going to be won on the front end, it is going to be won on the back end,” Landivar argued. “As you start investing in AI capabilities on the front end that help you make better promises, my advice is that you also invest in the underlying infrastructure that helps you keep that.”
